Obsessive-compulsive disorder (OCD) is a common psychiatric illness that often emerges in childhood and causes life-long disability in over 50% of patients. Psychological theory suggests that OCD symptoms are driven by difficulty disengaging affect from simple tasks (e.g. washing hands, locking a door) due to excessive anxiety about performance errors. Cognitive behavioral therapy (CBT), the gold standard treatment for OCD, repeatedly exposes patients to OC-relevant error cues during task performance until this anxiety habituates. While CBT is more effective in pediatric than adult samples, patients from both age groups are usually left with residual symptoms, highlighting the need for better treatments and raising the possibility that developmentally sensitive treatment may optimize outcomes for individual patients. Brain stimulation and/or cognitive training have the potential to augment CBT, but knowledge of the neural mechanisms of CBT is needed before these strategies can be appropriately targeted. Prior neuroimaging work in adolescent and adult OCD show common and distinct alterations of neural substrate for error-processing, even when OCD symptoms are not directly provoked. Specifically, adolescents and adults exhibit exaggerated anterior insula (aIns) activity, while adolescents show posterior medial frontal cortex (pMFC) hyperactivity that associates with lower symptom severity, and adults show atypical aIns - vmPFC connectivity that normalized with CBT. These findings point to developmentally sensitive alterations of error processing in the aIns, pMFC and vmPFC that may need to be targeted to enhance recovery and may need to be targeted differently at different stages of development. Based on these pilot data, CBT could work by increasing pMFC activity in youth, or by restoring the normal relationship between aIns and vmPFC in adults. In distinction to prior neuroimaging research, CBT will be studied in adolescent and adult patients, and in comparison to an active, non-CBT condition to control for non-specific effects of psychotherapy. Sixty adolescent (age 13-17) and 60 adult (age 25 - 45) patients, both with childhood onset OCD, will be randomized to either state-of-the art CBT for OCD, or stress management training (SMT), an active control therapy with minimal effects on OCD symptoms. Using a validated, incentivized performance monitoring task (Incentive Flanker Task, IFT) and resting state functional MRI, the following aims will be addressed: Aim 1) Demonstrate neural changes associated with CBT treatment, and Aim 2) Evaluate the effects of development on mechanisms of CBT. This knowledge will elucidate developmentally specific neural targets at which to direct brain stimulation therapy with transcranial magnetic stimulation and/or cognitive training to augment CBT for OCD.